Incorrect element index when accessing matrix in `cbuffer`

Description
DXC is generating incorrect element index when accessing matrix in a cbuffer.

Steps to Reproduce
https://godbolt.org/z/W1f1G1MYG
Compile with: dxc -T ps_6_6 -E main

cbuffer CB0 : register(b0) {
    float a;
    float b;
    float c;
    float1x1 m;
};
 
float4 main() : SV_Target
{
    float x = m._11;
    float y = a;
    float z = b;
    return float4(x, a, b, 0);
}

The generated code only accesses the at elements at indices 0 and 1. When accessing m._11 is it using the index 0 which is the index of a.

define void @main() {
  %CB0_cbuffer = call %dx.types.Handle @dx.op.createHandleFromBinding(i32 217, %dx.types.ResBind { i32 0, i32 0, i32 0, i8 2 }, i32 0, i1 false)  ; CreateHandleFromBinding(bind,index,nonUniformIndex)
  %1 = call %dx.types.Handle @dx.op.annotateHandle(i32 216, %dx.types.Handle %CB0_cbuffer, %dx.types.ResourceProperties { i32 13, i32 16 })  ; AnnotateHandle(res,props)  resource: CBuffer
  %2 = call %dx.types.CBufRet.f32 @dx.op.cbufferLoadLegacy.f32(i32 59, %dx.types.Handle %1, i32 0) ; line:24 col:15  ; CBufferLoadLegacy(handle,regIndex)
  %3 = extractvalue %dx.types.CBufRet.f32 %2, 0 ; line:24 col:15
  %4 = extractvalue %dx.types.CBufRet.f32 %2, 1 ; line:27 col:25
  call void @dx.op.storeOutput.f32(i32 5, i32 0, i32 0, i8 0, float %3) ; line:27 col:5  ; StoreOutput(outputSigId,rowIndex,colIndex,value)
  call void @dx.op.storeOutput.f32(i32 5, i32 0, i32 0, i8 1, float %3) ; line:27 col:5  ; StoreOutput(outputSigId,rowIndex,colIndex,value)
  call void @dx.op.storeOutput.f32(i32 5, i32 0, i32 0, i8 2, float %4) ; line:27 col:5  ; StoreOutput(outputSigId,rowIndex,colIndex,value)
  call void @dx.op.storeOutput.f32(i32 5, i32 0, i32 0, i8 3, float 0.000000e+00) ; line:27 col:5  ; StoreOutput(outputSigId,rowIndex,colIndex,value)
  ret void ; line:27 col:5
}

Actual Behavior
The m._11 value has element offset 3.

Environment

• DXC main as of 3/24/2025

Start by compiling the Godbolt reproducer with dxc -T ps_6_6 -E main and inspect the generated DXIL around cbufferLoadLegacy.f32. Trace how the m._11 access is assigned its element offset, then verify that the generated access uses offset 3 rather than the offset for a and that the other shown accesses remain correct.